Data Centre Sensitivity Analysis
Executive Summary
Key Takeaways
- ✓ Sensitivity analysis flexes one driver at a time, holding all others constant, to rank which individual assumptions most affect a model output, distinct from scenario analysis, which moves correlated drivers together.
- ✓ Occupancy, achieved price per kW, power cost, and PUE are typically the highest-impact individual drivers in a data centre model and should each be tested with its own sensitivity range.
- ✓ Sensitivity output should be ranked by magnitude of impact on the chosen output metric (revenue, EBITDA, or debt service coverage), a tornado-style presentation, so a reviewer can immediately see which assumption matters most.
- ✓ Sensitivity analysis identifies which individual assumption most affects an output; scenario analysis tests whether a coherent combination of assumptions produces an acceptable or unacceptable outcome. Both are necessary and neither substitutes for the other.
Objective¶
This guide sets out how to build driver-by-driver sensitivity analysis for a data centre financial model within Data Centre Financial Modelling.
Sensitivity Versus Scenario Analysis¶
Sensitivity analysis flexes one driver at a time, holding all other assumptions constant, to isolate and rank each individual driver's impact on a model output. This is distinct from Data Centre Scenario Analysis, which moves multiple correlated drivers together to test a coherent alternative future. Both serve necessary but different diagnostic purposes: sensitivity ranks individual assumption importance, while scenario testing evaluates whether a coherent combination of changes produces an acceptable outcome.
Highest-Impact Drivers¶
Occupancy, achieved price per kW, power cost, and PUE typically rank among the highest-impact individual drivers in a data centre model. Occupancy and price directly scale revenue; power cost and PUE directly scale one of the largest operating cost categories. Each should be tested with its own defined sensitivity range reflecting a realistic variation for that specific assumption, not a uniform percentage flex applied indiscriminately across all drivers. See Data Centre Financial KPIs for the broader KPI set these sensitivities test.
Presenting Sensitivity Output¶
Sensitivity output should be ranked by magnitude of impact on the chosen output metric, revenue, EBITDA, or debt service coverage, typically presented as a tornado chart or ranked table, so a reviewer can immediately see which individual assumption the output is most sensitive to without needing to separately interpret each variable's effect.
Using Sensitivity Analysis to Prioritise Review Effort¶
Because sensitivity analysis identifies which individual assumption matters most to the model's output, it should also inform where review and due diligence effort is concentrated: an assumption ranking highest in sensitivity impact warrants the most rigorous sourcing, documentation, and independent testing, while a low-impact assumption warrants proportionately less scrutiny.
Common Construction Pitfalls¶
Uniform percentage flex applied to all drivers regardless of realistic variation. Produces a sensitivity ranking that does not reflect each driver's actual plausible range.
Sensitivity analysis presented as a substitute for scenario analysis. Fails to test whether a coherent combination of changes produces an acceptable overall outcome.
Sensitivity output not ranked by magnitude. Forces a reviewer to interpret each variable's effect individually rather than immediately seeing which assumption matters most.
Recommended Practices¶
- Test each driver's sensitivity range individually, reflecting its own realistic variation.
- Rank sensitivity output by magnitude of impact on the chosen output metric.
- Use sensitivity ranking to prioritise review and due diligence effort.
- Run sensitivity and scenario analysis together, treating neither as a substitute for the other.
Continue Reading¶
Related Pillars¶
Related Technical Guides¶
Related Glossary¶
How OXXON tests thisRun a free structural check with FMAE
Frequently Asked Questions
How does sensitivity analysis differ from scenario analysis?
Sensitivity analysis flexes one driver at a time, holding all other assumptions constant, to isolate and rank each individual driver's impact on a model output. Scenario analysis instead moves multiple correlated drivers together to test a coherent alternative future. The two serve different diagnostic purposes and neither substitutes for the other.
Which drivers typically have the highest sensitivity impact in a data centre model?
Occupancy, achieved price per kW, power cost, and PUE typically rank among the highest-impact individual drivers, since occupancy and price directly scale revenue while power cost and PUE directly scale one of the largest operating cost categories.
How should sensitivity output be presented?
Ranked by magnitude of impact on the chosen output metric, revenue, EBITDA, or debt service coverage, typically as a tornado chart or ranked table, so a reviewer can immediately see which individual assumption the output is most sensitive to, without needing to interpret each variable's effect separately.
Why are both sensitivity and scenario analysis necessary?
Because sensitivity analysis identifies which individual assumption matters most, information useful for prioritising which assumptions warrant the most rigorous sourcing and review, while scenario analysis tests whether a coherent, correlated combination of assumptions produces an acceptable or unacceptable overall outcome, a question sensitivity analysis alone cannot answer.
Related Articles
Data Centre Financial Modelling
Data centre financial modelling is the discipline of modelling a data centre operator's revenue, cost, and capital structure from its capacity-denominated drivers, power, space, and cooling capacity, rack density, and tenant contract structure, rather than the generic market-price and headcount-growth drivers used in most corporate models, or the pure occupancy-and-lease-term drivers of conventional commercial real estate. This page is the hub for the Knowledge Centre's data centre financial modelling content: how colocation, hyperscale, and enterprise business models each require a distinct model architecture, how rack revenue and occupancy are decomposed into their separable underlying drivers, and how capacity planning and financial KPIs tie the model together, as this domain expands to cover operations, revenue, investment, and governance practice across the sector.
Data Centre Scenario Analysis
Data centre scenario analysis tests a model against structurally coherent alternative futures, correlated combinations of occupancy, pricing, power cost, and tenant concentration outcomes, rather than flexing a single driver in isolation. This guide sets out how to construct upside, base, and downside scenarios that move related drivers together consistently, and the sector-specific scenario dimensions, demand shift, power cost shock, and tenant concentration stress, most relevant to this business.
Data Centre Financial KPIs
A data centre financial model should track a defined set of KPIs spanning scale, utilisation, retention, pricing, efficiency, and profitability, since no single metric captures operating performance on its own. This guide sets out the core KPI set, MW under management, utilisation rate, churn, revenue per kW, power usage effectiveness, and EBITDA per MW, and how to interpret each correctly alongside the others rather than in isolation.
Power Usage Effectiveness (PUE)
Power usage effectiveness (PUE) is calculated as total facility power divided by critical IT load power, with a value approaching 1.0 indicating that nearly all power consumed is delivered to IT equipment rather than lost to cooling, power distribution, and other non-IT overhead. PUE is the standard industry measure of data centre power efficiency, and because power is typically one of the largest operating cost categories, a facility's PUE directly drives its power cost per unit of billable capacity and, in turn, its profitability.